This proposal responds to RFA DA-03-007, "The Impact of Child Psychopathology and Childhood Interventions on Subsequent Drug Use". It addresses the following questions from the RFA: 1. Which childhood psychiatric and behavioral conditions render individuals at greater risk for Substance Abuse Disorders (SUDs)? 2. How do childhood disorders or conditions render individuals vulnerable to later drug use problems and disorders? 3. What role do family factors play in child psychopathology and subsequent SUDs? 4. What role do environmental factors and traumatic events play in child psychopathology and subsequent SUDs? 5. Is it the severity of impairment associated with disorder, rather than the simple presence of a disorder, that is associated with increased risk for SUDs (dimensional approach)? 6. Does early mental health treatment reduce later risk for later SUDs? 7. Do some interventions unintentionally increase vulnerability to later SUDs? We propose to form a Consortium for Collaborative Data Analysis, made up of sites with longitudinal, epidemiologic data sets with data about both early psychiatric disorder and substance use and later SUDs. The aims are (1) to work together on generating publications from each site on topics relevant to the goals of the RFA; (2) to work together on common ways of handling issues such as handling missing data, models for longitudinal analysis, operationalizing common risk factors such as poverty or family stress, and diagnostic algorithms; (3) using a variant of standard meta-analytic methods which we call Multi-Site Secondary Analysis (MSSA), to work on combined analyses. MSSA avoids the 2 main limitations of standard meta-analysis: It does not rely only on published papers, which may present only a fraction of the data collected in each study, and which depends on investigators' decision to publish on a particular topic; and it is not subject to "publication bias": the tendency to publish only "significant" findings. MSSA controls for the different sample sizes, weighting schemes, and numbers of data waves of the different studies. Ten data sets comprising between 3 and 20 data waves on over 15,000 children have been assembled for the Consortium. Investigators will work together in weekly conference calls and 2 face-to-face meetings. Products will include a Supplement to the journal "Drug and Alcohol Dependence".